* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download CHAPTER 10 Mathematical Induction
Survey
Document related concepts
List of important publications in mathematics wikipedia , lookup
Philosophy of mathematics wikipedia , lookup
Turing's proof wikipedia , lookup
Foundations of mathematics wikipedia , lookup
Elementary mathematics wikipedia , lookup
Georg Cantor's first set theory article wikipedia , lookup
Peano axioms wikipedia , lookup
Collatz conjecture wikipedia , lookup
Fermat's Last Theorem wikipedia , lookup
Wiles's proof of Fermat's Last Theorem wikipedia , lookup
Fundamental theorem of algebra wikipedia , lookup
Brouwer–Hilbert controversy wikipedia , lookup
Four color theorem wikipedia , lookup
Transcript
CHAPTER 10 Mathematical Induction his chapter explains a powerful proof technique called mathematical induction (or just induction for short). To motivate the discussion, let’s first examine the kinds of statements that induction is used to prove. Consider the following statement. T Conjecture. The sum of the first n odd natural numbers equals n2 . The following table illustrates what this conjecture says. Each row is headed by a natural number n, followed by the sum of the first n odd natural numbers, followed by n2 . n sum of the first n odd natural numbers n2 1 2 3 4 5 .. . n .. . 1= ..................................... 1+3 = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1+3+5 = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1+3+5+7 = . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1+3+5+7+9 = . . . . . . . . . . . . . . . . . . . . . . . . .. . 1 + 3 + 5 + 7 + 9 + 11 + · · · + (2 n − 1) = . . . . . . . .. . 1 4 9 16 25 .. . n2 .. . Note that in the first five lines of the table, the sum of the first n odd numbers really does add up to n2 . Notice also that these first five lines indicate that the nth odd natural number (the last number in each sum) is 2n − 1. (For instance, when n = 2, the second odd natural number is 2 · 2 − 1 = 3; when n = 3, the third odd natural number is 2 · 3 − 1 = 5, etc.) The table raises a question. Does the sum 1 + 3 + 5 + 7 +· · ·+ (2 n − 1) really always equal n2 ? In other words, is the conjecture true? Let’s rephrase this as follows. For each natural number n (i.e., for each line of the table), we have a statement S n , as follows: 155 S 1 : 1 = 12 S 2 : 1 + 3 = 22 S 3 : 1 + 3 + 5 = 32 .. . S n : 1 + 3 + 5 + 7 + · · · + (2 n − 1) = n2 .. . Our question is: Are all of these statements true? Mathematical induction is designed to answer just this kind of question. It is used when we have a set of statements S1 , S2 , S3 , . . . , S n , . . ., and we need to prove that they are all true. The method is really quite simple. To visualize it, think of the statements as dominoes, lined up in a row. Imagine you can prove the first statement S1 , and symbolize this as domino S1 being knocked down. Additionally, imagine that you can prove that any statement S k being true (falling) forces the next statement S k+1 to be true (to fall). Then S1 falls, and knocks down S2 . Next S2 falls and knocks down S3 , then S3 knocks down S4 , and so on. The inescapable conclusion is that all the statements are knocked down (proved true). The Simple Idea Behind Mathematical Induction S6 ··· Sk Sk+1 Sk+2 Sk+3 Sk+4 · · · ··· Sk Sk+1 Sk+2 Sk+3 Sk+4 · · · Sk+2 Sk+3 Sk+4 · · · S k+ S5 S4 S k+ S3 S2 S k+ S1 Statements are lined up like dominoes. S1 S3 S2 S5 S4 S6 (1) Suppose the first statement falls (i.e. is proved true); S6 S5 S4 S3 S2 S1 ··· Sk Sk+ 1 (2) Suppose the k th falling always causes the ( k + 1) th to fall; 3 2 1 Then all must fall (i.e. all statements are proved true). Sk S6 S5 S4 S3 S2 S1 ··· ··· Mathematical Induction 156 This picture gives our outline for proof by mathematical induction. Proposition Outline for Proof by Induction The statements S1 , S2 , S3 , S4 , . . . are all true. Proof. (Induction) (1) Prove that the first statement S1 is true. (2) Given any integer k ≥ 1, prove that the statement S k ⇒ S k+1 is true. It follows by mathematical induction that every S n is true. ■ In this setup, the first step (1) is called the basis step. Because S1 is usually a very simple statement, the basis step is often quite easy to do. The second step (2) is called the inductive step. In the inductive step direct proof is most often used to prove S k ⇒ S k+1 , so this step is usually carried out by assuming S k is true and showing this forces S k+1 to be true. The assumption that S k is true is called the inductive hypothesis. Now let’s apply this technique to our original conjecture that the sum of the first n odd natural numbers equals n2 . Our goal is to show that for each n ∈ N, the statement S n : 1 + 3 + 5 + 7 + · · · + (2n − 1) = n2 is true. Before getting started, observe that S k is obtained from S n by plugging k in for n. Thus S k is the statement S k : 1 + 3 + 5 + 7 +· · ·+ (2k − 1) = k2 . Also, we get S k+1 by plugging in k + 1 for n, so that S k+1 : 1 + 3 + 5 + 7 +· · ·+ (2( k + 1) − 1) = (k + 1)2 . Proposition If n ∈ N, then 1 + 3 + 5 + 7 + · · · + (2 n − 1) = n2 . Proof. We will prove this with mathematical induction. (1) Observe that if n = 1, this statement is 1 = 12 , which is obviously true. (2) We must now prove S k ⇒ S k+1 for any k ≥ 1. That is, we must show that if 1+3+5+7+· · ·+(2 k −1) = k2 , then 1+3+5+7+· · ·+(2( k +1)−1) = (k +1)2 . We use direct proof. Suppose 1 + 3 + 5 + 7 + · · · + (2 k − 1) = k2 . Then 1 + 3 + 5 + 7 + · · · · · · · · · · · · · · · + (2( k + 1) − 1) = 1 + 3 + 5 + 7 + · · · + (2 k − 1) + (2( k + 1) − 1) = ¡ ¢ 1 + 3 + 5 + 7 + · · · + (2 k − 1) + (2( k + 1) − 1) = k2 + (2( k + 1) − 1) = k2 + 2 k + 1 = ( k + 1)2 . Thus 1 + 3 + 5 + 7 + · · · + (2( k + 1) − 1) = (k + 1)2 . This proves that S k ⇒ S k+1 . It follows by induction that 1 + 3 + 5 + 7 + · · · + (2n − 1) = n2 for every n ∈ N. ■ 157 In induction proofs it is usually the case that the first statement S 1 is indexed by the natural number 1, but this need not always be so. Depending on the problem, the first statement could be S0 , or S m for any other integer m. In the next example the statements are S0 , S1 , S2 , S3 , . . . The same outline is used, except that the basis step verifies S0 , not S1 . Proposition If n is a non-negative integer, then 5 | ( n5 − n). Proof. We will prove this with mathematical induction. Observe that the first non-negative integer is 0, so the basis step involves n = 0. (1) If n = 0, this statement is 5 | (05 − 0) or 5 | 0, which is obviously true. (2) Let k ≥ 0. We need to prove that if 5 | ( k5 − k), then 5 | ((k + 1)5 − ( k + 1)). We use direct proof. Suppose 5 | ( k5 − k). Thus k5 − k = 5a for some a ∈ Z. Observe that ( k + 1)5 − ( k + 1) = k5 + 5 k4 + 10 k3 + 10 k2 + 5 k + 1 − k − 1 = ( k5 − k) + 5 k4 + 10 k3 + 10 k2 + 5 k = 5a + 5 k4 + 10 k3 + 10 k2 + 5 k = 5(a + k4 + 2 k3 + 2 k2 + k). This shows ( k + 1)5 − (k + 1) is an integer multiple of 5, so 5 | ((k + 1)5 − ( k + 1)). We have now shown that 5 | (k5 − k) implies 5 | ((k + 1)5 − (k + 1)). It follows by induction that 5 | (n5 − n) for all non-negative integers n. ■ As noted, induction is used to prove statements of the form ∀ n ∈ N, S n . But notice the outline does not work for statements of form ∀ n ∈ Z, S n (where n is in Z, not N). The reason is that if you are trying to prove ∀ n ∈ Z, S n by induction, and you’ve shown S 1 is true and S k ⇒ S k+1 , then it only follows from this that S n is true for n ≥ 1. You haven’t proved that any of the statements S0 , S−1 , S−2 , . . . are true. If you ever want to prove ∀ n ∈ Z, S n by induction, you have to show that some S a is true and S k ⇒ S k+1 and S k ⇒ S k−1 . Unfortunately, the term mathematical induction is sometimes confused with inductive reasoning, that is, the process of reaching the conclusion that something is likely to be true based on prior observations of similar circumstances. Please note that that mathematical induction, as introduced here, is a rigorous method that proves statements with absolute certainty. Mathematical Induction 158 To round out this section, we present four additional induction proofs. Proposition If n ∈ Z and n ≥ 0, then n X i · i ! = ( n + 1)! − 1. i =0 Proof. We will prove this with mathematical induction. (1) If n = 0, this statement is 0 X i · i ! = (0 + 1)! − 1. i =0 Since the left-hand side is 0 · 0! = 0, and the right-hand side is 1! − 1 = 0, P the equation 0i=0 i · i ! = (0 + 1)! − 1 holds, as both sides are zero. (2) Consider any integer k ≥ 0. We must show that S k implies S k+1 . That is, we must show that k X i · i ! = ( k + 1)! − 1 i =0 implies kX +1 i · i ! = (( k + 1) + 1)! − 1. i =0 k X We use direct proof. Suppose i · i ! = ( k + 1)! − 1. Observe that i =0 kX +1 à k X i · i! = i =0 ! i · i ! + ( k + 1)( k + 1)! i =0 = ³ ´ ( k + 1)! − 1 + ( k + 1)( k + 1)! = ( k + 1)! + ( k + 1)( k + 1)! − 1 ¡ ¢ = 1 + ( k + 1) ( k + 1)! − 1 = ( k + 2)( k + 1)! − 1 = ( k + 2)! − 1 = (( k + 1) + 1)! − 1. Therefore kX +1 i · i ! = (( k + 1) + 1)! − 1. i =0 It follows by induction that n X i =0 i · i ! = ( n + 1)! − 1 for every integer n ≥ 0. ■ 159 The next example illustrates a trick that is occasionally useful. You know that you can add equal quantities to both sides of an equation without violating equality. But don’t forget that you can add unequal quantities to both sides of an inequality, as long as the quantity added to the bigger side is bigger than the quantity added to the smaller side. For example, if x ≤ y and a ≤ b, then x + a ≤ y + b. Similarly, if x ≤ y and b is positive, then x ≤ y + b. This oft-forgotten fact is used in the next proof. For each n ∈ N, it follows that 2n ≤ 2n+1 − 2n−1 − 1. Proposition Proof. We will prove this with mathematical induction. (1) If n = 1, this statement is 21 ≤ 21+1 − 21−1 − 1, which simplifies to 2 ≤ 4 − 1 − 1, which is obviously true. (2) Suppose k ≥ 1. We need to show that 2k ≤ 2k+1 − 2k−1 − 1 implies 2k+1 ≤ 2(k+1)+1 −2(k+1)−1 −1. We use direct proof. Suppose 2k ≤ 2k+1 −2k−1 −1, and reason as follows: 2k ≤ 2k+1 − 2k−1 − 1 2(2k ) ≤ 2(2k+1 − 2k−1 − 1) 2 k+1 2 k+1 2 k+1 2 k+1 ≤ 2 k+2 −2 −2 ≤ 2 k+2 − 2k − 2 + 1 ≤ 2 k+2 (add 1 to the bigger side) k −2 −1 ( k+1)+1 ≤ 2 (multiply both sides by 2) k − 2(k+1)−1 − 1. It follows by induction that 2n ≤ 2n+1 − 2n−1 − 1 for each n ∈ N. ■ We next prove that if n ∈ N, then the inequality (1 + x)n ≥ 1 + nx holds for all x ∈ R with x > −1. Thus we will need to prove that the statement S n : (1 + x)n ≥ 1 + nx for every x ∈ R with x > −1 is true for every natural number n. This is (only) slightly different from our other examples, which proved statements of the form ∀n ∈ N, P ( n), where P ( n) is a statement about the number n. This time we are proving something of form ∀ n ∈ N, P ( n, x), where the statement P (n, x) involves not only n, but also a second variable x. (For the record, the inequality (1 + x)n ≥ 1 + nx is known as Bernoulli’s inequality.) Mathematical Induction 160 Proposition If n ∈ N, then (1 + x)n ≥ 1 + nx for all x ∈ R with x > −1. Proof. We will prove this with mathematical induction. (1) For the basis step, notice that when n = 1 the statement is (1 + x)1 ≥ 1 + 1 · x , and this is true because both sides equal 1 + x. (2) Assume that for some k ≥ 1, the statement (1 + x)k ≥ 1 + kx is true for all x ∈ R with x > −1. From this we need to prove (1 + x)k+1 ≥ 1 + ( k + 1) x. Now, 1 + x is positive because x > −1, so we can multiply both sides of (1 + x)k ≥ 1 + kx by (1 + x) without changing the direction of the ≥. (1 + x)k (1 + x) ≥ (1 + kx)(1 + x) (1 + x)k+1 ≥ 1 + x + kx + kx2 (1 + x)k+1 ≥ 1 + ( k + 1) x + kx2 The above term kx2 is positive, so removing it from the right-hand side will only make that side smaller. Thus we get (1 + x)k+1 ≥ 1 + ( k + 1) x. ■ Next, an example where the basis step involves more than routine checking. (It will be used later, so it is numbered for reference.) Proposition 10.1 Suppose a 1 , a 2 , . . . , a n are n integers, where n ≥ 2. If p is prime and p | (a 1 · a 2 · a 3 · · · a n ), then p | a i for at least one of the a i . Proof. The proof is induction on n. (1) The basis step involves n = 2. Let p be prime and suppose p | (a 1 a 2 ). We need to show that p | a 1 or p | a 2 , or equivalently, if p - a 1 , then p | a 2 . Thus suppose p - a 1 . Since p is prime, it follows that gcd( p, a 1 ) = 1. By Proposition 7.1 (on page 126), there are integers k and ` for which 1 = pk + a 1 `. Multiplying this by a 2 gives a 2 = pka 2 + a 1 a 2 `. As we are assuming that p divides a 1 a 2 , it is clear that p divides the expression pka 2 + a 1 a 2 ` on the right; hence p | a 2 . We’ve now proved that if p | (a 1 a 2 ), then p | a 1 or p | a 2 . This completes the basis step. (2) Suppose that k ≥ 2, and p | (a 1 · a 2 · · · a k ) implies then p | a i for some a i . ¡ ¢ Now let p | (a 1 · a 2 · · · a k · a k+1 ). Then p | (a 1 · a 2 · · · a k ) · a k+1 . By what we proved in the basis step, it follows that p | (a 1 · a 2 · · · a k ) or p | a k+1 . This and the inductive hypothesis imply that p divides one of the a i . ■ Please test your understanding now by working a few exercises. Proof by Strong Induction 161 10.1 Proof by Strong Induction This section describes a useful variation on induction. Occasionally it happens in induction proofs that it is difficult to show that S k forces S k+1 to be true. Instead you may find that you need to use the fact that some “lower” statements S m (with m < k) force S k+1 to be true. For these situations you can use a slight variant of induction called strong induction. Strong induction works just like regular induction, except that in Step (2) instead of assuming S k is true and showing this forces S k+1 to be true, we assume that all the statements S1 , S2 , . . . , S k are true and show this forces S k+1 to be true. The idea is that if it always happens that the first k dominoes falling makes the (k + 1)th domino fall, then all the dominoes must fall. Here is the outline. Outline for Proof by Strong Induction Proposition The statements S1 , S2 , S3 , S4 , . . . are all true. Proof. (Strong induction) (1) Prove the first statement S1 . (Or the first several S n .) (2) Given any integer k ≥ 1, prove (S1 ∧ S2 ∧ S3 ∧ · · · ∧ S k ) ⇒ S k+1 . ■ Strong induction can be useful in situations where assuming S k is true does not neatly lend itself to forcing S k+1 to be true. You might be better served by showing some other statement (S k−1 or S k−2 for instance) forces S k to be true. Strong induction says you are allowed to use any (or all) of the statements S1 , S2 , . . . , S k to prove S k+1 . As our first example of strong induction, we are going to prove that 12 | ( n4 − n2 ) for any n ∈ N. But first, let’s look at how regular induction would be problematic. In regular induction we would start by showing 12 | ( n4 − n2 ) is true for n = 1. This part is easy because it reduces to 12 | 0, which is clearly true. Next we would assume that 12 | ( k4 − k2 ) and try to show this implies 12 | ((k + 1)4 − (k + 1)2 ). Now, 12 | ( k4 − k2 ) means k4 − k2 = 12a for some a ∈ Z. Next we use this to try to show (k + 1)4 − ( k + 1)2 = 12b for some integer b. Working out ( k + 1)4 − ( k + 1)2 , we get ( k + 1)4 − ( k + 1)2 = ( k4 + 4 k3 + 6 k2 + 4 k + 1) − ( k2 + 2 k + 1) = ( k4 − k2 ) + 4 k3 + 6 k2 + 6 k = 12a + 4 k3 + 6 k2 + 6 k. At this point we’re stuck because we can’t factor out a 12. Now let’s see how strong induction can get us out of this bind. Mathematical Induction 162 Strong induction involves assuming each of statements S1 , S2 , . . . , S k is true, and showing that this forces S k+1 to be true. In particular, if S1 through S k are true, then certainly S k−5 is true, provided that 1 ≤ k − 5 < k. The idea is then to show S k−5 ⇒ S k+1 instead of S k ⇒ S k+1 . For this to make sense, our basis step must involve checking that S1 , S2 , S3 , S4 , S5 , S6 are all true. Once this is established, S k−5 ⇒ S k+1 will imply that the other S k are all true. For example, if k = 6, then S k−5 ⇒ S k+1 is S 1 ⇒ S 7 , so S 7 is true; for k = 7, then S k−5 ⇒ S k+1 is S2 ⇒ S8 , so S8 is true, etc. Proposition If n ∈ N, then 12 | ( n4 − n2 ). Proof. We will prove this with strong induction. (1) First note that the statement is true for the first six positive integers: If n = 1, 12 divides n4 − n2 = 14 − 12 = 0. If n = 2, 12 divides n4 − n2 = 24 − 22 = 12. If n = 3, 12 divides n4 − n2 = 34 − 32 = 72. If n = 4, 12 divides n4 − n2 = 44 − 42 = 240. If n = 5, 12 divides n4 − n2 = 54 − 52 = 600. If n = 6, 12 divides n4 − n2 = 64 − 62 = 1260. (2) Let k ≥ 6 and assume 12 | (m4 − m2 ) for 1 ≤ m ≤ k. (That is, assume ¡ ¢ statements S1 , S2 , . . . , S k are all true.) We must show 12 | ( k + 1)4 − ( k + 1)2 . (That is, we must show that S k+1 is true.) Since S k−5 is true, we have 12 | (( k − 5)4 − ( k − 5)2 ). For simplicity, let’s set m = k − 5, so we know 12 | ( m4 − m2 ), meaning ( k + 1)4 − ( k + 1)2 m4 − m2 = 12a for some integer a. Observe that: = ( m + 6)4 − ( m + 6)2 = m4 + 24 m3 + 216 m2 + 864 m + 1296 − ( m2 + 12 m + 36) = ( m4 − m2 ) + 24 m3 + 216 m2 + 852 m + 1260 = 12a + 24 m3 + 216 m2 + 852 m + 1260 ¡ ¢ = 12 a + 2 m3 + 18 m2 + 71 m + 105 . As (a + 2m3 + 18 m2 + 71 m + 105) is an integer, we get 12 | ((k + 1)4 − (k + 1)2 ). This shows by strong induction that 12 | ( n4 − n2 ) for every n ∈ N. ■ Proof by Strong Induction 163 Our next example involves mathematical objects called graphs. In mathematics, the word graph is used in two contexts. One context involves the graphs of equations and functions from algebra and calculus. In the other context, a graph is a configuration consisting of points (called vertices) and edges which are lines connecting the vertices. Following are some pictures of graphs. Let’s agree that all of our graphs will be in “one piece,” that is, you can travel from any vertex of a graph to any other vertex by traversing a route of edges from one vertex to the other. v0 v1 v4 v2 v3 Figure 10.1. Examples of Graphs A cycle in a graph is a sequence of distinct edges in the graph that form a route that ends where it began. For example, the graph on the far left of Figure 10.1 has a cycle that starts at vertex v1 , then goes to v2 , then to v3 , then v4 and finally back to its starting point v1 . You can find cycles in both of the graphs on the left, but the two graphs on the right do not have cycles. There is a special name for a graph that has no cycles; it is called a tree. Thus the two graphs on the right of Figure 10.1 are trees, but the two graphs on the left are not trees. Figure 10.2. A tree Note that the trees in Figure 10.1 both have one fewer edge than vertex. The tree on the far right has 5 vertices and 4 edges. The one next to it has 6 vertices and 5 edges. Draw any tree; you will find that if it has n vertices, then it has n − 1 edges. We now prove that this is always true. Mathematical Induction 164 Proposition If a tree has n vertices, then it has n − 1 edges. Proof. Notice that this theorem asserts that for any n ∈ N, the following statement is true: S n : A tree with n vertices has n − 1 edges. We use strong induction to prove this. (1) Observe that if a tree has n = 1 vertex then it has no edges. Thus it has n − 1 = 0 edges, so the theorem is true when n = 1. (2) Now take an integer k ≥ 1. We must show (S1 ∧ S2 ∧ · · · ∧ S k ) ⇒ S k+1 . In words, we must show that if it is true that any tree with m vertices has m − 1 edges, where 1 ≤ m ≤ k, then any tree with k + 1 vertices has ( k + 1) − 1 = k edges. We will use direct proof. Suppose that for each integer m with 1 ≤ m ≤ k, any tree with m vertices has m − 1 edges. Now let T be a tree with k + 1 vertices. Single out an edge of T and label it e, as illustrated below. ··· ··· ··· T ··· 1 T e ··· ··· T2 ··· ··· Now remove the edge e from T , but leave the two endpoints of e. This leaves two smaller trees that we call T1 and T2 . Let’s say T1 has x vertices and T2 has y vertices. As each of these two smaller trees has fewer than k + 1 vertices, our inductive hypothesis guarantees that T1 has x − 1 edges, and T2 has y − 1 edges. Think about our original tree T . It has x + y vertices. It has x − 1 edges that belong to T1 and y − 1 edges that belong to T2 , plus it has the additional edge e that belongs to neither T1 nor T2 . Thus, all together, the number of edges that T has is ( x − 1) + ( y − 1) + 1 = ( x + y) − 1. In other words, T has one fewer edges than it has vertices. Thus it has (k + 1) − 1 = k edges. It follows by strong induction that a tree with n vertices has n − 1 edges. ■ Notice that it was absolutely essential that we used strong induction in the above proof because the two trees T1 and T2 will not both have k vertices. At least one will have fewer than k vertices. Thus the statement S k is not enough to imply S k+1 . We need to use the assumption that S m will be true whenever m ≤ k, and strong induction allows us to do this. Proof by Smallest Counterexample 165 10.2 Proof by Smallest Counterexample This section introduces yet another proof technique, called proof by smallest counterexample. It is a hybrid of induction and proof by contradiction. It has the nice feature that it leads you straight to a contradiction. It is therefore more “automatic” than the proof by contradiction that was introduced in Chapter 6. Here is the outline: Outline for Proof by Smallest Counterexample Proposition The statements S1 , S2 , S3 , S4 , . . . are all true. Proof. (Smallest counterexample) (1) Check that the first statement S1 is true. (2) For the sake of contradiction, suppose not every S n is true. (3) Let k > 1 be the smallest integer for which S k is false. (4) Then S k−1 is true and S k is false. Use this to get a contradiction. ■ Notice that this setup leads you to a point where S k−1 is true and S k is false. It is here, where true and false collide, that you will find a contradiction. Let’s do an example. Proposition If n ∈ N, then 4 | (5n − 1). Proof. We use proof by smallest counterexample. (We will number the steps to match the outline, but that is not usually done in practice.) (1) If n = 1, then the statement is 4 | (51 − 1), or 4 | 4, which is true. (2) For sake of contradiction, suppose it’s not true that 4 | (5n − 1) for all n. (3) Let k > 1 be the smallest integer for which 4 - (5k − 1). (4) Then 4 | (5k−1 − 1), so there is an integer a for which 5k−1 − 1 = 4a. Then: 5k−1 − 1 = 4a 5(5k−1 − 1) = 5 · 4a 5k − 5 = 20a 5k − 1 = 20a + 4 5k − 1 = 4(5a + 1) This means 4 | (5k − 1), a contradiction, because 4 - (5k − 1) in Step 3. Thus, we were wrong in Step 2 to assume that it is untrue that 4 | (5n − 1) for every n. Therefore 4 | (5n − 1) is true for every n. ■ 166 Mathematical Induction We next prove the fundamental theorem of arithmetic, which says any integer greater than 1 has a unique prime factorization. For example, 12 factors into primes as 12 = 2 · 2 · 3, and moreover any factorization of 12 into primes uses exactly the primes 2, 2 and 3. Our proof combines the techniques of induction, cases, minimum counterexample and the idea of uniqueness of existence outlined at the end of Section 7.3. We dignify this fundamental result with the label of “Theorem.” Theorem 10.1 (Fundamental Theorem of Arithmetic) Any integer n > 1 has a unique prime factorization. That is, if n = p 1 · p 2 · p 3 · · · p k and n = a 1 · a 2 · a 3 · · · a ` are two prime factorizations of n, then k = `, and the primes p i and a i are the same, except that they may be in a different order. Proof. Suppose n > 1. We first use strong induction to show that n has a prime factorization. For the basis step, if n = 2, it is prime, so it is already its own prime factorization. Let n ≥ 2 and assume every integer between 2 and n (inclusive) has a prime factorization. Consider n + 1. If it is prime, then it is its own prime factorization. If it is not prime, then it factors as n + 1 = ab with a, b > 1. Because a and b are both less than n + 1 they have prime factorizations a = p 1 · p 2 · p 3 · · · p k and b = p01 · p02 · p03 · · · p0` . Then n + 1 = ab = ( p 1 · p 2 · p 3 · · · p k )( p01 · p02 · p03 · · · p0` ) is a prime factorization of n + 1. This competes the proof by strong induction that every integer greater than 1 has a prime factorization. Next we use proof by smallest counterexample to prove that the prime factorization of any n ≥ 2 is unique. If n = 2, then n clearly has only one prime factorization, namely itself. Assume for the sake of contradiction that there is an n > 2 that has different prime factorizations n = p 1 · p 2 · p 3 · · · p k and n = a 1 · a 2 · a 3 · · · a ` . Assume n is the smallest number with this property. From n = p 1 · p 2 · p 3 · · · p k , we see that p 1 | n, so p 1 | (a 1 · a 2 · a 3 · · · a ` ). By Proposition 10.1 (page 160), p 1 divides one of the primes a i . As a i is prime, we have p 1 = a i . Dividing n = p 1 · p 2 · p 3 · · · p k = a 1 · a 2 · a 3 · · · a ` by p 1 = a i yields p 2 · p 3 · · · p k = a 1 · a 2 · a 3 · · · a i−1 · a i+1 · · · a ` . These two factorizations are different, because the two prime factorizations of n were different. (Remember: the primes p 1 and a i are equal, so the difference appears in the remaining factors, displayed above.) But also the above number p 2 · p 3 · · · p k is smaller than n, and this contradicts the fact that n was the smallest number with two different prime factorizations. ■ Fibonacci Numbers 167 One word of warning about proof by smallest counterexample. In proofs in other textbooks or in mathematical papers, it often happens that the writer doesn’t tell you up front that proof by smallest counterexample is being used. Instead, you will have to read through the proof to glean from context that this technique is being used. In fact, the same warning applies to all of our proof techniques. If you continue with mathematics, you will gradually gain through experience the ability to analyze a proof and understand exactly what approach is being used when it is not stated explicitly. Frustrations await you, but do not be discouraged by them. Frustration is a natural part of anything that’s worth doing. 10.3 Fibonacci Numbers Leonardo Pisano, now known as Fibonacci, was a mathematician born around 1175 in what is now Italy. His most significant work was a book Liber Abaci, which is recognized as a catalyst in medieval Europe’s slow transition from Roman numbers to the Hindu-Arabic number system. But he is best known today for a number sequence that he described in his book and that bears his name. The Fibonacci sequence is 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, . . . The numbers that appear in this sequence are called Fibonacci numbers. The first two numbers are 1 and 1, and thereafter any entry is the sum of the previous two entries. For example 3 + 5 = 8, and 5 + 8 = 13, etc. We denote the nth term of this sequence as F n . Thus F1 = 1, F2 = 1, F3 = 2, F4 = 3, F7 = 13 and so on. Notice that the Fibonacci Sequence is entirely determined by the rules F1 = 1, F2 = 1, and F n = F n−1 + F n−2 . We introduce Fibonacci’s sequence here partly because it is something everyone should know about, but also because it is a great source of induction problems. This sequence, which appears with surprising frequency in nature, is filled with mysterious patterns and hidden structures. Some of these structures will be revealed to you in the examples and exercises. We emphasize that the condition F n = F n−1 + F n−2 (or equivalently F n+1 = F n + F n−1 ) is the perfect setup for induction. It suggests that we can determine something about F n by looking at earlier terms of the sequence. In using induction to prove something about the Fibonacci sequence, you should expect to use the equation F n = F n−1 + F n−2 somewhere. For our first example we will prove that F n2+1 − F n+1 F n − F n2 = (−1)n for any natural number n. For example, if n = 5 we have F62 − F6 F5 − F52 = 82 − 8 · 5 − 52 = 64 − 40 − 25 = −1 = (−1)5 . Mathematical Induction 168 Proposition The Fibonacci sequence obeys F n2+1 − F n+1 F n − F n2 = (−1)n . Proof. We will prove this with mathematical induction. (1) If n = 1 we have F n2+1 − F n+1 F n − F n2 = F22 − F2 F1 − F12 = 12 − 1 · 1 − 12 = −1 = (−1)1 = (−1)n , so indeed F n2+1 − F n+1 F n − F n2 = (−1)n is true when n = 1. (2) Take any integer k ≥ 1. We must show that if F k2+1 − F k+1 F k − F k2 = (−1)k , then F k2+2 − F k+2 F k+1 − F k2+1 = (−1)k+1 . We use direct proof. Suppose F k2+1 − F k+1 F k − F k2 = (−1)k . Now we are going to carefully work out the expression F k2+2 − F k+2 F k+1 − F k2+1 and show that it really does equal (−1)k+1 . In so doing, we will use the fact that F k+2 = F k+1 + F k . F k2+2 − F k+2 F k+1 − F k2+1 = (F k+1 + F k )2 − (F k+1 + F k )F k+1 − F k2+1 = F k2+1 + 2F k+1 F k + F k2 − F k2+1 − F k F k+1 − F k2+1 = −F k2+1 + F k+1 F k + F k2 = −(F k2+1 − F k+1 F k − F k2 ) = −(−1)k (inductive hypothesis) = (−1)1 (−1)k = (−1)k+1 Therefore F k2+2 − F k+2 F k+1 − F k2+1 = (−1)k+1 . It follows by induction that F n2+1 − F n+1 F n − F n2 = (−1)n for every n ∈ N. ■ Let’s pause for a moment and think about what the result we just proved means. Dividing both sides of F n2+1 − F n+1 F n − F n2 = (−1)n by F n2 gives µ F n+1 Fn ¶2 − F n+1 (−1)n −1 = . Fn F n2 For large values of n, the right-hand side is very close to zero, and the left-hand side is F n+1 /F n plugged into the polynomial x2 − x − 1. Thus, as n increases, the ratio F n+1 /F n approaches a p root of x2 − x − 1 = 0. By the quadratic formula, the roots of x2 − x − 1 are 1±2 5 . As F n+1 /F n > 1, this ratio must be approaching the positive root p 1+ 5 2 . Therefore p F n+1 1 + 5 lim = . n→∞ F n 2 (10.1) p For a quick spot check, note that F13 /F12 ≈ 1.618025, while 1+2 5 ≈ 1.618033. Even for the small value n = 12, the numbers match to four decimal places. Fibonacci Numbers 169 p The number Φ = 1+2 5 is sometimes called the golden ratio, and there has been much speculation about its occurrence in nature as well as in classical art and architecture. One theory holds that the Parthenon and the Great Pyramids of Egypt were designed in accordance with this number. But we are here concerned with things that can be proved. We close by observing how the Fibonacci sequence in many ways resembles a geometric sequence. Recall that a geometric sequence with first term a and common ratio r has the form a, ar, ar 2 , ar 3 , ar 4 , ar 5 , ar 6 , ar 7 , ar 8 , . . . where any term is obtained by multiplying the previous term by r . In general its nth term is G n = ar n , and G n+1 /G n = r . Equation (10.1) tells us that F n+1 /F n ≈ Φ. Thus even though it is not a geometric sequence, the Fibonacci sequence tends to behave like a geometric sequence with common ratio Φ, and the further “out” you go, the higher the resemblance. Exercises for Chapter 10 Prove the following statements with either induction, strong induction or proof by smallest counterexample. n2 + n . 2 n( n + 1)(2 n + 1) For every integer n ∈ N, it follows that 12 + 22 + 32 + 42 + · · · + n2 = . 6 n2 ( n + 1)2 For every integer n ∈ N, it follows that 13 + 23 + 33 + 43 + · · · + n3 = . 4 n( n + 1)( n + 2) . If n ∈ N, then 1 · 2 + 2 · 3 + 3 · 4 + 4 · 5 + · · · + n(n + 1) = 3 If n ∈ N, then 21 + 22 + 23 + · · · + 2n = 2n+1 − 2. n X For every natural number n, it follows that (8 i − 5) = 4 n2 − n. 1. For every integer n ∈ N, it follows that 1 + 2 + 3 + 4 + · · · + n = 2. 3. 4. 5. 6. i =1 7. If n ∈ N, then 1 · 3 + 2 · 4 + 3 · 5 + 4 · 6 + · · · + n(n + 2) = n( n + 1)(2 n + 7) . 6 1 2 3 n 1 + + +···+ = 1− 2! 3! 4! ( n + 1)! ( n + 1)! 9. For any integer n ≥ 0, it follows that 24 | (52n − 1). 8. If n ∈ N, then 10. For any integer n ≥ 0, it follows that 3 | (52n − 1). 11. For any integer n ≥ 0, it follows that 3 | (n3 + 5 n + 6). 12. For any integer n ≥ 0, it follows that 9 | (43n + 8). Mathematical Induction 170 13. For any integer n ≥ 0, it follows that 6 | (n3 − n). 14. Suppose a ∈ Z. Prove that 5 | 2n a implies 5 | a for any n ∈ N. 1 1 1 1 1 1 + + + +···+ = 1− . 1·2 2·3 3·4 4·5 n( n + 1) n+1 16. For every natural number n, it follows that 2n + 1 ≤ 3n . 15. If n ∈ N, then 17. Suppose A 1 , A 2 , . . . A n are sets in some universal set U , and n ≥ 2. Prove that A1 ∩ A2 ∩ · · · ∩ A n = A1 ∪ A2 ∪ · · · ∪ A n . 18. Suppose A 1 , A 2 , . . . A n are sets in some universal set U , and n ≥ 2. Prove that A1 ∪ A2 ∪ · · · ∪ A n = A1 ∩ A2 ∩ · · · ∩ A n . 1 1 1 1 1 + + +···+ 2 ≤ 2− . 1 4 9 n n 20. Prove that (1 + 2 + 3 + · · · + n)2 = 13 + 23 + 33 + · · · + n3 for every n ∈ N. 1 1 1 1 1 1 1 n 21. If n ∈ N, then + + + + + · · · + n + ≥ 1+ . 1 2 3 4 5 2 − 1 2n 2 (Note: This problem asserts that the sum of the first 2n terms of the harmonic series is at least 1 + n/2. It thus implies that the harmonic series diverges.) ¶µ ¶µ ¶µ ¶ µ ¶ µ 1 1 1 1 1 1 1 22. If n ∈ N, then 1 − 1− 1− 1− · · · 1 − n ≥ + n+1 . 2 4 8 16 2 4 2 19. Prove that 23. Use mathematical induction to prove the binomial theorem (Theorem 3.1 on page 80). You may find that you need Equation (3.2) on page 78. 24. Prove that n X ¡ ¢ k nk = n2n−1 for each natural number n. k=1 25. Concerning the Fibonacci sequence, prove that F1 + F2 + F3 + F4 + . . . + F n = F n+2 − 1. 26. Concerning the Fibonacci sequence, prove that n X F k2 = F n F n+1 . k=1 27. Concerning the Fibonacci sequence, prove that F1 + F3 + F5 + F7 + . . . + F2n−1 = F2n . 28. Concerning the Fibonacci sequence, prove that F2 + F4 + F6 + F8 + . . . + F2n = F2n+1 − 1. 29. In this problem n ∈ N and F n is the nth Fibonacci number. Prove that ¡ n ¢ ¡ n−1 ¢ ¡ n−2 ¢ ¡ n−3 ¢ ¡0¢ 0 + 1 + 2 + 3 + · · · + n = F n+1 . (For example, ¡6¢ 0 + ¡5¢ 1 + ¡4¢ 2 + ¡3¢ 3 + ¡2¢ 4 + ¡1¢ 5 + ¡0¢ 6 = 1+5+6+1+0+0+0 = 13 = F6+1 .) 30. Here F n is the nth Fibonacci number. Prove that ³ Fn = 31. Prove that n ¡ ¢ X k r k=0 = p ´n ³ p ´n 1+ 5 − 1−2 5 2 p 5 . ¡ n+1 ¢ r +1 , where 1 ≤ r ≤ n. 32. Prove that the number of n-digit binary numbers that have no consecutive 1’s is the Fibonacci number F n+2 . For example, for n = 2 there are three such Fibonacci Numbers 171 numbers (00, 01, and 10), and 3 = F2+2 = F4 . Also, for n = 3 there are five such numbers (000, 001, 010, 100, 101), and 5 = F3+2 = F5 . 33. Suppose n (infinitely long) straight lines lie on a plane in such a way that no two of the lines are parallel, and no three of the lines intersect at a single 2 point. Show that this arrangement divides the plane into n +2n+2 regions. 3n+1 − 3 for every n ∈ N. 2 ¡ ¢ 35. Prove that if n, k ∈ N, and n is even and k is odd, then nk is even. 34. Prove that 31 + 32 + 33 + 34 + · · · + 3n = 36. Prove that if n = 2k − 1 for some k ∈ N, then every entry in the nth row of Pascal’s triangle is odd. The remaining odd-numbered exercises below are not solved in the back of the book. 37. Prove that if m, n ∈ N, then n P k k=0 ¡ m+ k ¢ m =n ¡m+n+1¢ ¡m+n+1¢ m+1 − m+2 . 38. Prove that if n is a positive integer, then 39. Prove that if n is a positive integer, then 40. Prove that 41. Prove that p ¡ ¢¡ P m k=0 m ¡ P k=0 ¡n¢2 ¡2n¢ ¡n¢2 ¡n¢2 0 + 1 + 2 +···+ n = n . ¡n+0¢ ¡n+1¢ ¡n+2¢ ¡n+k¢ ¡n+k+1¢ . 0 + 1 + 2 +···+ k = k ¡n¢2 n ¢ ¡ m+ n¢ p− k = p for positive integers m, n and p. m¢¡ n ¢ ¡ m+ n¢ k p + k = m+ p for positive integers m, n and p. k